论文部分内容阅读
提出了一种QoS约束的多目标优化的网格工作流调度算法ISPEA2,该算法基于表达结构丰富的AGWL网格工作流模型,在SPEA2算法中引入约束检测对网格工作流调度问题进行优化,克服了当前网格工作流调度算法中大多只考虑DAG结构的网格工作流、涉及QoS参数较少及将多QoS参数聚合成一个单目标函数进行优化调度的缺陷,可供决策者根据用户的实际需求从产生的Pareto优化解集中选择最终的满意解。通过与采用原SPEA2设计的网格工作流调度算法OSPEA2的比较,表明ISPEA2算法所获得的Pareto优化解集都是满足QoS约束的非支配解,且获得了更优的平均结果。
A QoS constrained multi-objective optimization workflow scheduling algorithm ISPEA2 is proposed. This algorithm is based on the structure-rich workflow model of AGWL grid, introduces the constraint detection in SPEA2 algorithm to optimize the workflow scheduling problem, It overcomes the grid workflow that most of the current grid workflow scheduling algorithms only consider the DAG structure, involves less QoS parameters and aggregates multiple QoS parameters into a single objective function for optimal scheduling. It can be used by decision makers based on the user’s The actual demand selects the final satisfactory solution from the set of Pareto optimal solutions generated. The comparison with OSPEA2, which uses the original SPEA2 design workflow algorithm, shows that the Pareto optimal solution set obtained by ISPEA2 algorithm is a non-dominated solution satisfying the QoS constraint and obtains a better average result.